Ascertaining first-degree relatives in adult-onset cancer research: The genetic epidemiology of lung cancer study, 1999-2003 Free

B60

Successful recruitment of family members related to an affected individual is essential for many family-based study designs. In family-based case-control studies designed to assess correlations between genetic variants in candidate genes, use of unaffected family controls may increase control response rates and reduce the effect of population stratification. Feasibility of recruiting parents and siblings of probands with adult-onset disease has not been documented, particularly among African Americans.
 We contacted via telephone 379 individuals with early-onset (age <50) lung cancer (270 white, 109 African Americans) identified through the Metropolitan Detroit Surveillance, Epidemiology and End Results program. We collected information on family structure and also attempted to interview both parents and one sibling for each proband. We examined completeness of data collected for first degree relatives by race, sex and whether the information was obtained through the lung cancer proband or a proxy using chi-square tests. We also examined the accuracy of data collected by parent and sibling self-report versus proband report using Pearson’s correlation coefficients.
 White individuals with lung cancer were more likely to have a living mother and living father than their African American counterparts (both p-values=0.02). Among living parents, 46.1% of white mothers, 31.2% of African American mothers, 42.0% of white fathers, and 7.1% of African American fathers participated in the interview and donated a biospecimen. Concordance of reported histories of lung disease (COPD, allergies, asthma, chronic bronchitis, and tuberculosis) between cancer proband interview and personal interview of the parents was high, with percent concordant ranging from 80-100%. Among all mothers and white fathers, correlations between proband and parent reports of number of years of smoking, usual number of cigarettes per day, and pack years of smoking were also high (range 0.64-0.99). Concordance could not be assessed for African American fathers due to sample size (n=5).
 Nearly all individuals with lung cancer had at least one living full sibling over the age of 18 who would have been eligible to participate in the study (93.4% of white cases and 95% of African American cases). White and African American probands were equally likely to have a sibling participate in the interview (53.9% and 47.8%, respectively; p=0.16), and siblings of white and African American probands were equally likely to provide a biospecimen (p=0.18). Similar to parental data, concordance rates between self-reported and cancer proband interview data were high.
 Enrolling parents and siblings of individuals with adult-onset disease into population-based studies is feasible, although response rates vary by race and relationship to the proband. Among our population, the concordance of health and smoking related reports were high. However, our estimates may be higher than they would be for the average family relationship given that the relatives who participated in this study were identified through the proband and may be relatives with whom probands have high levels of communication regarding health-related issues. Further work to examine why individuals choose to participate in family-based studies is warranted.